trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 490,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.04,
      "learning_rate": 7.959183673469388e-06,
      "loss": 0.4344,
      "step": 100
    },
    {
      "epoch": 4.08,
      "learning_rate": 5.918367346938776e-06,
      "loss": 0.171,
      "step": 200
    },
    {
      "epoch": 6.12,
      "learning_rate": 3.877551020408164e-06,
      "loss": 0.1436,
      "step": 300
    },
    {
      "epoch": 8.16,
      "learning_rate": 1.8367346938775512e-06,
      "loss": 0.1312,
      "step": 400
    }
  ],
  "max_steps": 490,
  "num_train_epochs": 10,
  "total_flos": 2032907146629120.0,
  "trial_name": null,
  "trial_params": null
}